Merge pull request #32279 from vespa-engine/bratseth/gram-root

Make the gram root settable
vespa-engine · Aug 28, 2024 · 4abdba0 · 4abdba0
2 parents feb9ff2 + d0e2471
commit 4abdba0
Show file tree

Hide file tree

Showing 2 changed files with 45 additions and 5 deletions.
diff --git a/container-search/src/main/java/com/yahoo/search/querytransform/NGramSearcher.java b/container-search/src/main/java/com/yahoo/search/querytransform/NGramSearcher.java
@@ -15,11 +15,16 @@
 import com.yahoo.prelude.query.CompositeItem;
 import com.yahoo.prelude.query.HasIndexItem;
 import com.yahoo.prelude.query.Item;
+import com.yahoo.prelude.query.NearItem;
+import com.yahoo.prelude.query.ONearItem;
+import com.yahoo.prelude.query.OrItem;
 import com.yahoo.prelude.query.PhraseItem;
 import com.yahoo.prelude.query.SegmentItem;
 import com.yahoo.prelude.query.Substring;
 import com.yahoo.prelude.query.TermItem;
+import com.yahoo.prelude.query.WeakAndItem;
 import com.yahoo.prelude.query.WordItem;
+import com.yahoo.processing.request.CompoundName;
 import com.yahoo.search.Query;
 import com.yahoo.search.Result;
 import com.yahoo.search.Searcher;
@@ -44,6 +49,8 @@
 @After(JUNIPER_TAG_REPLACING)
 public class NGramSearcher extends Searcher {
 
+    private static final CompoundName gramMatch = CompoundName.from("gram.match");
+
     private final GramSplitter gramSplitter;
 
     private final CharacterClasses characterClasses;
@@ -151,7 +158,16 @@ protected CompositeItem createGramRoot(HasIndexItem term, Query query) {
 
     /** Creates the root of the query subtree without access to the term being replaced. */
     protected CompositeItem createGramRoot(Query query) {
-        return new AndItem();
+        return switch (query.properties().getString(gramMatch, "all")) {
+            case "all" -> new AndItem();
+            case "any" -> new OrItem();
+            case "weakAnd" -> new WeakAndItem();
+            case "phrase" -> new PhraseItem();
+            case "near" -> new NearItem();
+            case "onear" -> new ONearItem();
+            default -> throw new IllegalArgumentException("Invalid gram.match value '" + query.properties().getString(gramMatch) +
+                                                          "'. Must be 'all', 'any', 'weakAnd', 'phrase', 'near' or 'onear'");
+        };
     }
 
     private void replaceItemByGrams(Item item, Item grams, int indexInParent) {

diff --git a/...iner-search/src/test/java/com/yahoo/search/querytransform/test/NGramSearcherTestCase.java b/...iner-search/src/test/java/com/yahoo/search/querytransform/test/NGramSearcherTestCase.java
@@ -173,6 +173,30 @@ void testNGramRewritingShortOnly() {
         assertEquals("WEAKAND(100) gram3:en", q.getModel().getQueryTree().toString());
     }
 
+    @Test
+    void testSettingGramMatching() {
+        assertEquals("WEAKAND(100) (AND gram2:en gram2:ng)", search("?query=gram2:eng"));
+        assertEquals("WEAKAND(100) (AND gram2:en gram2:ng)", search("?query=gram2:eng&gram.match=all"));
+        assertEquals("WEAKAND(100) (OR gram2:en gram2:ng)", search("?query=gram2:eng&gram.match=any"));
+        assertEquals("WEAKAND(100) (WEAKAND(100) gram2:en gram2:ng)", search("?query=gram2:eng&gram.match=weakAnd"));
+        assertEquals("WEAKAND(100) gram2:\"en ng\"", search("?query=gram2:eng&gram.match=phrase"));
+        assertEquals("WEAKAND(100) (NEAR(2) gram2:en gram2:ng)", search("?query=gram2:eng&gram.match=near"));
+        assertEquals("WEAKAND(100) (ONEAR(2) gram2:en gram2:ng)", search("?query=gram2:eng&gram.match=onear"));
+        try {
+            search("?query=gram2:eng&gram.match=invalid");
+            fail("Expected exception");
+        } catch (IllegalArgumentException e) {
+            assertEquals("Invalid gram.match value 'invalid'. Must be 'all', 'any', 'weakAnd', 'phrase', 'near' or 'onear'",
+                         e.getMessage());
+        }
+    }
+
+    String search(String query) {
+        Query q = new Query(query);
+        createExecution().search(q);
+        return q.getModel().getQueryTree().toString();
+    }
+
     @Test
     void testNGramRewritingShortInMixes() {
         Query q = new Query("?query=test:a+gram3:en");
@@ -335,19 +359,19 @@ void testNGramRecombining() {
         Hit h1 = r.hits().get("hit1");
         assertEquals("Should be untouched,\u001feven if containing \u001f",
                      h1.getField("test").toString());
-        assertTrue(h1.getField("test") instanceof String);
+        assertInstanceOf(String.class, h1.getField("test"));
 
         assertEquals("Blue red Ed A", h1.getField("gram2").toString());
-        assertTrue(h1.getField("gram2") instanceof XMLString);
+        assertInstanceOf(XMLString.class, h1.getField("gram2"));
 
         assertEquals("Blue red ed a\u001f",
                      h1.getField("gram3").toString(),
                      "Separators on borders work");
-        assertTrue(h1.getField("gram3") instanceof String);
+        assertInstanceOf(String.class, h1.getField("gram3"));
 
         Hit h2 = r.hits().get("hit2");
         assertEquals("katt  i...morgen", h2.getField("gram3").toString());
-        assertTrue(h2.getField("gram3") instanceof JSONString);
+        assertInstanceOf(JSONString.class, h2.getField("gram3"));
 
         Hit h3 = r.hits().get("hit3");
         assertEquals("\u001ffin\u001f \u001fen\u001f \u001fa\u001f", h3.getField("gram2").toString());